Failed to complete upgrade 8.7 -> 9.1

My VPS was running Centos 8 and I successfully upgrade to RockyLinux version 8. And after that, I did install RockyLinux 9, and here is where all my problems began.

My last success was updating the kernel (6.0.10-1.el9.elrepo.x86_64)

I have 395 packages that I didn’t get updated on the system. I tried many solutions but until now nothing. This is URL contains all packages with warning dependency

[root@vmi731405 rpm]# dnf update -y
Last metadata expiration check: 1:45:01 ago on Mon 28 Nov 2022 02:49:05 PM EST.
Error:
 Problem 1: cannot install the best update candidate for package GConf2-3.2.6-22.el8.x86_64
  - nothing provides libc.so.6(GLIBC_2.34)(64bit) needed by GConf2-3.2.6-34.el9.x86_64
 Problem 2: cannot install the best update candidate for package alsa-lib-1.2.7.2-1.el8.x86_64
  - nothing provides libc.so.6(GLIBC_2.34)(64bit) needed by alsa-lib-1.2.7.2-1.el9.x86_64
  - nothing provides libm.so.6(GLIBC_2.29)(64bit) needed by alsa-lib-1.2.7.2-1.el9.x86_64
 Problem 3: cannot install the best update candidate for package alsa-tools-firmware-1.1.6-1.el8.x86_64
  - nothing provides libc.so.6(GLIBC_2.34)(64bit) needed by alsa-tools-firmware-1.2.2-6.el9.x86_64

[root@vmi731405 rpm]# dnf -y --releasever=9 --allowerasing --setopt=deltarpm=false distro-sync --nobest --skip-broken
Last metadata expiration check: 1:39:54 ago on Mon 28 Nov 2022 02:49:05 PM EST.
Error:
 Problem: The operation would result in removing the following protected packages: dnf, kernel-ml-core, sudo, systems

[root@vmi731405 rpm]#  rpm  -qa | grep kernel
kernel-ml-modules-6.0.10-1.el9.elrepo.x86_64
kernel-ml-6.0.10-1.el9.elrepo.x86_64
kernel-srpm-macros-1.0-11.el9.noarch
kernel-ml-core-6.0.10-1.el9.elrepo.x86_64

[root@vmi731405 rpm]# ls -la /etc/yum.repos.d/
drwxr-xr-x.   3 root root  4096 Nov 28 14:48 .
drwxr-xr-x. 122 root root 12288 Nov 28 14:55 ..
-rw-r--r--    1 root root  1946 Jul 10 07:36 elrepo.repo
-rw-r--r--    1 root root  1552 Aug 24 04:37 epel-testing.repo
-rw-r--r--    1 root root  1453 Aug 24 04:37 epel.repo
drwxr-xr-x    4 root root  4096 Nov 28 09:56 old3
-rw-r--r--    1 root root  6586 Oct 20 11:50 rocky-addons.repo
-rw-r--r--    1 root root   409 Oct 20 11:50 rocky-devel.repo
-rw-r--r--    1 root root  1990 Nov 28 10:01 rocky-extras.repo
-rw-r--r--    1 root root  2379 Oct 20 11:50 rocky-extras.repo.rpmnew
-rw-r--r--    1 root root  2015 Nov 28 09:57 rocky.repo
-rw-r--r--    1 root root  3405 Oct 20 11:50 rocky.repo.rpmnew

research URLs:

https://bugzilla.redhat.com/show_bug.cgi?id=1768468
https://forums.rockylinux.org/t/rocky-linux-8-upgrade-to-rocky-linux-9/6629/3
https://computingforgeeks.com/upgrade-to-rocky-linux-9-from-rocky-linux-8/
https://linoxide.com/upgrade-to-latest-linux-kernel-on-almalinux-8-or-rocky/
https://www.thegeekdiary.com/cannot-install-the-best-update-candidate-for-package-error-on-running-dnf-update/

How did you try to perform the upgrade from 8.7 to 9.1? What are the steps that you took to try this upgrade?

Note: Upgrades are not supported and there are no supported methods provided from us.

Rocky 8 and Rocky 9 are two distinct distros. Granted, they are somewhat similar.

Do you have backups of configs and user data? That is, can you reinstall, deploy configs, and restore data? Being able to do so is useful even when one does not attempt anything unsupported.

Is the VPS a VM? Was there possiblity to create a snapshot before you started? Snapshot is a coarse-grained backup that some use to spawn multiple VM instances.

You don’t have to go back to CentOS; restoring from Rocky 8.7 snapshot is fine, if possible.

1 Like

I only have Centos’s snapshot. I forgot to create it before upgrading Rocky from 8 to 9.

In general, the “user data”; documents, pictures, web sites, databases, should be kept separate to the o/s, and snapshots are not real backups anyway, so normally you would just re-install clean Rocky 9.1 and then restore the data.

If the CentOS snapshot has the data you need, you could try restoring it, and then copy the “user data” back into the new o/s.

Your o/s is unsupported in multiple ways, the prominent error is to do with glibc, you could run

rpm -q glibc
rpm --verify glibc

you might also be able to copy the “user data” from the existing broken o/s, and restore it back into a new clean install.